Take Home Exercise 02

Analysing and Visualising the distribution of Airbnb listings in 2019 and 2021, and how Covid-19 affected the distribution.

Kwek Yi Chen https://example.com/kwekyichen
09-16-2021

1 Introduction

In recent years, the emergence of numerous peer-based business models has empowered individuals across the globe to become micro-entrepreneurs, earning money from their idle property and spare time. This phenomenon, entitled ‘the sharing economy’, has seen unprecedented growth in terms of user numbers, enabling new avenues of economic and social interaction (Sundararajan, 2016).

Since its inception in 2008, Airbnb has expanded into over 34,000 cities across 191 countries. Uncommonly for a sharing economy company, Airbnb transitioned into profitability in 2016, demonstrating proof-of-concept for the validity of sharing economy within the global market (Stone & Zaleski, 2017).

Interestingly, Singapore is one of the global city that has yet to legalise short-term rentals offered by platforms such as Airbnb (Read more at https://www.todayonline.com/singapore/short-term-home-sharing-remains-illegal-singapore-airbnb-disappointed). However, during my recent visit to Inside Airbnb (http://insideairbnb.com/about.html), an independent, non-commercial set of tools and data that allows anyone to explore how Airbnb is really being used in cities around the world, it came to my attention that there are data sets for Singapore.

This analysis consist of two sections: 1) Airbnb Distribution in 2019 2) Impact of COVID-19

2 Data

3 Setting Up the Environment

The following code chunk install the following packages:

Show code 
packages = c('maptools', 'sf', 'raster','spatstat', 'tmap', 'tidyr', 'readr')
for (p in packages){
  if(!require(p, character.only = T)){
    install.packages(p)
  }
  library(p,character.only = T)
}

4 Spatial Data Wrangling

4.1 Importing Geospatial Data

4.1.1 MrtLrt

4.1.1.1 Importing

The following code uses st_read function of sf package to import MrtLrt geospatial data sets.

Show code 
mrtlrt_sf <- st_read(dsn = "data/geospatial", layer="MRTLRTStnPtt")

Reading layer `MRTLRTStnPtt' from data source 
  `C:\kwekyichen\IS415_blog\_posts\2021-09-16-takehome-ex02\data\geospatial' 
  using driver `ESRI Shapefile'
Simple feature collection with 185 features and 3 fields
Geometry type: POINT
Dimension:     XY
Bounding box:  xmin: 6138.311 ymin: 27555.06 xmax: 45254.86 ymax: 47854.2
Projected CRS: SVY21

The MrtLrt data have 185 features with 3 fields. The projected CRS is SVY21. The geometry type is in point.

4.1.1.2 Check CRS

Next, we use st_crs function of sf to check the CRS.

Show code 
st_crs(mrtlrt_sf)

Coordinate Reference System:
  User input: SVY21 
  wkt:
PROJCRS["SVY21",
    BASEGEOGCRS["SVY21[WGS84]",
        DATUM["World Geodetic System 1984",
            ELLIPSOID["WGS 84",6378137,298.257223563,
                LENGTHUNIT["metre",1]],
            ID["EPSG",6326]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["Degree",0.0174532925199433]]],
    CONVERSION["unnamed",
        METHOD["Transverse Mercator",
            ID["EPSG",9807]],
        PARAMETER["Latitude of natural origin",1.36666666666667,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8801]],
        PARAMETER["Longitude of natural origin",103.833333333333,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8802]],
        PARAMETER["Scale factor at natural origin",1,
            SCALEUNIT["unity",1],
            ID["EPSG",8805]],
        PARAMETER["False easting",28001.642,
            LENGTHUNIT["metre",1],
            ID["EPSG",8806]],
        PARAMETER["False northing",38744.572,
            LENGTHUNIT["metre",1],
            ID["EPSG",8807]]],
    CS[Cartesian,2],
        AXIS["(E)",east,
            ORDER[1],
            LENGTHUNIT["metre",1,
                ID["EPSG",9001]]],
        AXIS["(N)",north,
            ORDER[2],
            LENGTHUNIT["metre",1,
                ID["EPSG",9001]]]]

From the result above, even though the mrtlrt_sf is projected in svy21, the last line indicates that the EPSG is 9001. It is wrong because the correct EPSG code for svy21 should be 3414.

4.1.1.3 Assign EPSG code to a simple feature data frame

The following code chunk will assign the correct EPSG code to mrtlrt_sf using st_set_crs function of sf package

Show code 
mrtlrt_sf <- st_set_crs(mrtlrt_sf, 3414)

Now that we have assign the correct EPSG code to mrtlrt_sf, we will check the CRS again using st_crs function of sf

4.1.1.4 Check CRS again

Show code 
st_crs(mrtlrt_sf)

Coordinate Reference System:
  User input: EPSG:3414 
  wkt:
PROJCRS["SVY21 / Singapore TM",
    BASEGEOGCRS["SVY21",
        DATUM["SVY21",
            ELLIPSOID["WGS 84",6378137,298.257223563,
                LENGTHUNIT["metre",1]]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["degree",0.0174532925199433]],
        ID["EPSG",4757]],
    CONVERSION["Singapore Transverse Mercator",
        METHOD["Transverse Mercator",
            ID["EPSG",9807]],
        PARAMETER["Latitude of natural origin",1.36666666666667,
            ANGLEUNIT["degree",0.0174532925199433],
            ID["EPSG",8801]],
        PARAMETER["Longitude of natural origin",103.833333333333,
            ANGLEUNIT["degree",0.0174532925199433],
            ID["EPSG",8802]],
        PARAMETER["Scale factor at natural origin",1,
            SCALEUNIT["unity",1],
            ID["EPSG",8805]],
        PARAMETER["False easting",28001.642,
            LENGTHUNIT["metre",1],
            ID["EPSG",8806]],
        PARAMETER["False northing",38744.572,
            LENGTHUNIT["metre",1],
            ID["EPSG",8807]]],
    CS[Cartesian,2],
        AXIS["northing (N)",north,
            ORDER[1],
            LENGTHUNIT["metre",1]],
        AXIS["easting (E)",east,
            ORDER[2],
            LENGTHUNIT["metre",1]],
    USAGE[
        SCOPE["Cadastre, engineering survey, topographic mapping."],
        AREA["Singapore - onshore and offshore."],
        BBOX[1.13,103.59,1.47,104.07]],
    ID["EPSG",3414]]

From the result above, we have successfully assign EPSG 3414 to mrtlrt_sf.

4.1.2 CostalOutline

4.1.2.1 Importing

The following code uses st_read function of sf package to import CostalOutline geospatial data sets.

Show code 
sg_sf <- st_read(dsn = "data/geospatial", layer="CostalOutline")

Reading layer `CostalOutline' from data source 
  `C:\kwekyichen\IS415_blog\_posts\2021-09-16-takehome-ex02\data\geospatial' 
  using driver `ESRI Shapefile'
Simple feature collection with 60 features and 4 fields
Geometry type: POLYGON
Dimension:     XY
Bounding box:  xmin: 2663.926 ymin: 16357.98 xmax: 56047.79 ymax: 50244.03
Projected CRS: SVY21

The CostalOutline data have 60 features with 4 fields. The projected CRS is SVY21. The geometry type is polygon.

4.1.2.2 Check CRS

Show code 
st_crs(sg_sf)

Coordinate Reference System:
  User input: SVY21 
  wkt:
PROJCRS["SVY21",
    BASEGEOGCRS["SVY21[WGS84]",
        DATUM["World Geodetic System 1984",
            ELLIPSOID["WGS 84",6378137,298.257223563,
                LENGTHUNIT["metre",1]],
            ID["EPSG",6326]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["Degree",0.0174532925199433]]],
    CONVERSION["unnamed",
        METHOD["Transverse Mercator",
            ID["EPSG",9807]],
        PARAMETER["Latitude of natural origin",1.36666666666667,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8801]],
        PARAMETER["Longitude of natural origin",103.833333333333,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8802]],
        PARAMETER["Scale factor at natural origin",1,
            SCALEUNIT["unity",1],
            ID["EPSG",8805]],
        PARAMETER["False easting",28001.642,
            LENGTHUNIT["metre",1],
            ID["EPSG",8806]],
        PARAMETER["False northing",38744.572,
            LENGTHUNIT["metre",1],
            ID["EPSG",8807]]],
    CS[Cartesian,2],
        AXIS["(E)",east,
            ORDER[1],
            LENGTHUNIT["metre",1,
                ID["EPSG",9001]]],
        AXIS["(N)",north,
            ORDER[2],
            LENGTHUNIT["metre",1,
                ID["EPSG",9001]]]]

Similarly, even though the sg_sf is projected in svy21, the last line indicates that the EPSG is 9001. It is wrong because the correct EPSG code for svy21 should be 3414.

4.1.2.3 Assign EPSG code to a simple feature data frame

The following code chunk will assign the correct EPSG code to sg_sf using st_set_crs function of sf package

Show code 
sg_sf <- st_set_crs(sg_sf, 3414)

Now that we have assign the correct EPSG code to sg_sf, we will check the CRS again using st_crs function of sf

4.1.2.4 Check CRS again

Show code 
st_crs(sg_sf)

Coordinate Reference System:
  User input: EPSG:3414 
  wkt:
PROJCRS["SVY21 / Singapore TM",
    BASEGEOGCRS["SVY21",
        DATUM["SVY21",
            ELLIPSOID["WGS 84",6378137,298.257223563,
                LENGTHUNIT["metre",1]]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["degree",0.0174532925199433]],
        ID["EPSG",4757]],
    CONVERSION["Singapore Transverse Mercator",
        METHOD["Transverse Mercator",
            ID["EPSG",9807]],
        PARAMETER["Latitude of natural origin",1.36666666666667,
            ANGLEUNIT["degree",0.0174532925199433],
            ID["EPSG",8801]],
        PARAMETER["Longitude of natural origin",103.833333333333,
            ANGLEUNIT["degree",0.0174532925199433],
            ID["EPSG",8802]],
        PARAMETER["Scale factor at natural origin",1,
            SCALEUNIT["unity",1],
            ID["EPSG",8805]],
        PARAMETER["False easting",28001.642,
            LENGTHUNIT["metre",1],
            ID["EPSG",8806]],
        PARAMETER["False northing",38744.572,
            LENGTHUNIT["metre",1],
            ID["EPSG",8807]]],
    CS[Cartesian,2],
        AXIS["northing (N)",north,
            ORDER[1],
            LENGTHUNIT["metre",1]],
        AXIS["easting (E)",east,
            ORDER[2],
            LENGTHUNIT["metre",1]],
    USAGE[
        SCOPE["Cadastre, engineering survey, topographic mapping."],
        AREA["Singapore - onshore and offshore."],
        BBOX[1.13,103.59,1.47,104.07]],
    ID["EPSG",3414]]

We have successfully set the EPSG to 3414 for sg_sf.

4.1.3 MPSZ

4.1.3.1 Importing

The following code uses st_read function of sf package to import MP14_SUBZONE_WEB_PL geospatial data sets.

Show code 
mpsz_sf <- st_read(dsn = "data/geospatial", layer = "MP14_SUBZONE_WEB_PL")

Reading layer `MP14_SUBZONE_WEB_PL' from data source 
  `C:\kwekyichen\IS415_blog\_posts\2021-09-16-takehome-ex02\data\geospatial' 
  using driver `ESRI Shapefile'
Simple feature collection with 323 features and 15 fields
Geometry type: MULTIPOLYGON
Dimension:     XY
Bounding box:  xmin: 2667.538 ymin: 15748.72 xmax: 56396.44 ymax: 50256.33
Projected CRS: SVY21

The mpsz data have 323 features with 15 fields. The projected CRS is SVY21. The geometry type is multipolygon.

4.1.3.2 Check CRS

Show code 
st_crs(mpsz_sf)

Coordinate Reference System:
  User input: SVY21 
  wkt:
PROJCRS["SVY21",
    BASEGEOGCRS["SVY21[WGS84]",
        DATUM["World Geodetic System 1984",
            ELLIPSOID["WGS 84",6378137,298.257223563,
                LENGTHUNIT["metre",1]],
            ID["EPSG",6326]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["Degree",0.0174532925199433]]],
    CONVERSION["unnamed",
        METHOD["Transverse Mercator",
            ID["EPSG",9807]],
        PARAMETER["Latitude of natural origin",1.36666666666667,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8801]],
        PARAMETER["Longitude of natural origin",103.833333333333,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8802]],
        PARAMETER["Scale factor at natural origin",1,
            SCALEUNIT["unity",1],
            ID["EPSG",8805]],
        PARAMETER["False easting",28001.642,
            LENGTHUNIT["metre",1],
            ID["EPSG",8806]],
        PARAMETER["False northing",38744.572,
            LENGTHUNIT["metre",1],
            ID["EPSG",8807]]],
    CS[Cartesian,2],
        AXIS["(E)",east,
            ORDER[1],
            LENGTHUNIT["metre",1,
                ID["EPSG",9001]]],
        AXIS["(N)",north,
            ORDER[2],
            LENGTHUNIT["metre",1,
                ID["EPSG",9001]]]]

Similarly, even though the mpsz_sf is projected in svy21, the last line indicates that the EPSG is 9001. It is wrong because the correct EPSG code for svy21 should be 3414.

4.1.3.3 Assign EPSG code to a simple feature data frame

The following code chunk will assign the correct EPSG code to mpsz_sf using st_set_crs function of sf package

Show code 
mpsz_sf <- st_set_crs(mpsz_sf, 3414)

Now that we have assign the correct EPSG code to mpsz_sf, we will check the CRS again using st_crs function of sf

4.1.3.4 Check CRS again

Show code 
st_crs(mpsz_sf)

Coordinate Reference System:
  User input: EPSG:3414 
  wkt:
PROJCRS["SVY21 / Singapore TM",
    BASEGEOGCRS["SVY21",
        DATUM["SVY21",
            ELLIPSOID["WGS 84",6378137,298.257223563,
                LENGTHUNIT["metre",1]]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["degree",0.0174532925199433]],
        ID["EPSG",4757]],
    CONVERSION["Singapore Transverse Mercator",
        METHOD["Transverse Mercator",
            ID["EPSG",9807]],
        PARAMETER["Latitude of natural origin",1.36666666666667,
            ANGLEUNIT["degree",0.0174532925199433],
            ID["EPSG",8801]],
        PARAMETER["Longitude of natural origin",103.833333333333,
            ANGLEUNIT["degree",0.0174532925199433],
            ID["EPSG",8802]],
        PARAMETER["Scale factor at natural origin",1,
            SCALEUNIT["unity",1],
            ID["EPSG",8805]],
        PARAMETER["False easting",28001.642,
            LENGTHUNIT["metre",1],
            ID["EPSG",8806]],
        PARAMETER["False northing",38744.572,
            LENGTHUNIT["metre",1],
            ID["EPSG",8807]]],
    CS[Cartesian,2],
        AXIS["northing (N)",north,
            ORDER[1],
            LENGTHUNIT["metre",1]],
        AXIS["easting (E)",east,
            ORDER[2],
            LENGTHUNIT["metre",1]],
    USAGE[
        SCOPE["Cadastre, engineering survey, topographic mapping."],
        AREA["Singapore - onshore and offshore."],
        BBOX[1.13,103.59,1.47,104.07]],
    ID["EPSG",3414]]

We have successfully assign ESPG 3414 to mpsz_sf.

4.2 Importing Aspatial Data

4.2.1 Hotel

4.2.1.1 Importing

The following code uses read_csv function of readr package to import Hotel aspatial data sets.

Show code 
hotels <- read_csv("data/aspatial/hotels.csv")

4.2.1.2 Examine the file

The following code uses list function of base R to examine the imported file

Show code 
list(hotels)

[[1]]
# A tibble: 422 x 9
   NAME     ADDRESSPOSTALCODE ADDRESSSTREETNAME HYPERLINK   TOTALROOMS
   <chr>                <dbl> <chr>             <chr>            <dbl>
 1 Jayleen~             59390 25 New Bridge Ro~ jayleencla~         20
 2 JEN Sin~            238858 277 Orchard Road~ singaporeo~        499
 3 JEN Sin~            249716 1A Cuscaden Road  singaporet~        565
 4 JH HOTEL            399041 8 LORONG 10 GEYL~ anguskt@gm~         42
 5 Ji Hote~            238485 11 PENANG LANE I~ <NA>                81
 6 Jin Don~            399007 33 Lorong 12 Gey~ <NA>                33
 7 Joyfor ~            389226 135 Geylang Road  joyfor.sg@~         17
 8 JW Marr~            189763 30 Beach Road     jw.sinjw.c~        634
 9 K Hotel             399199 44 LORONG 6 GEYL~ <NA>                56
10 Holiday~            423371 88 EAST COAST RO~ <NA>               451
# ... with 412 more rows, and 4 more variables: KEEPERNAME <chr>,
#   Lat <dbl>, Lng <dbl>, ICON_NAME <chr>

The data contains 422 rows and 9 columns. There are ‘Lat’ and ‘Lng’ column for coordinates. They are in decimal degree format, hence we will assume the data is wgs84.

We will next check if there are any NA value in ‘Lat’ and ‘Lng’ using is.na function of base R package.

Show code 
any(is.na(hotels$Lat))

[1] FALSE
Show code 
any(is.na(hotels$Lng))

[1] FALSE

We can confirmed that there are no NA values and we can proceed to create a simple feature data frame.

4.2.1.3 Create simple feature data frame from aspatial data

The following code 1. convert hotels data into simple feature data frame using st_as_sf feature of sf package. 2. ‘Lng’ and ‘Lat’ column are set as the coords arguments. 3. CRS were set as 4326 since we assumed that the data is wgs84 4. st_transform function of sf package then helps to transform the newly created simple feature data frame into svy21 5. tm_shape and tm_dots function of tmap plots the transformed hotels_sf

Show code 
hotels_sf <- st_as_sf(hotels,
                    coords = c("Lng", 
                               "Lat"),
                    crs=4326) %>%
  st_transform(crs = 3414)

tm_shape(hotels_sf)+
  tm_dots()

4.2.2 Tourism

4.2.2.1 Importing

The following code uses read_csv function of readr package to import Tourism aspatial data sets.

Show code 
tourism <- read_csv("data/aspatial/tourism.csv")

4.2.2.2 Examine the file

The following code uses list function of base R to examine the imported file

Show code 
list(tourism)

[[1]]
# A tibble: 107 x 17
   NAME   DESCRIPTION  ADDRESSSTREETNA~ HYPERLINK  PHOTOURL  URL_PATH 
   <chr>  <chr>        <chr>            <chr>      <chr>     <chr>    
 1 China~ Experience ~ 48 Pagoda Street http://ww~ www.your~ www.your~
 2 Thian~ Beautifully~ 158 Telok Ayer ~ http://ww~ www.your~ www.your~
 3 Euras~ This engagi~ 139 Ceylon Road  http://ww~ www.your~ www.your~
 4 Shoph~ With many b~ <NA>             <NA>       www.your~ www.your~
 5 Capit~ The queen o~ 11 Stamford Road http://ww~ www.your~ www.your~
 6 Espla~ This boldly~ 1 Esplanade Dri~ http://ww~ www.your~ www.your~
 7 Garde~ The lush an~ 18 Marina Garde~ http://ww~ www.your~ www.your~
 8 The I~ There’s mor~ 180-226 Depot R~ http://ww~ <NA>      www.your~
 9 PARKR~ PARKROYAL o~ 3 Upper Pickeri~ http://ww~ www.your~ www.your~
10 Lasal~ How can som~ 1 McNally Street https://w~ www.your~ www.your~
# ... with 97 more rows, and 11 more variables: IMAGE_ALT_TEXT <chr>,
#   PHOTOCREDITS <chr>, LASTMODIFIED <dttm>, LATITUDE <dbl>,
#   LONGTITUDE <dbl>, META_DESCRIPTION <chr>, OPENING_HOURS <chr>,
#   Lat <dbl>, Lng <dbl>, ICON_NAME <chr>, ADDRESSPOSTALCODE <dbl>

The tourism data contains 107 rows and 17 columns. There are ‘LATITUDE’ and ‘LONGTITUDE’ column, as well as ‘Lat’ and ‘Lng’ coluumn for coordinates. We will be using ‘LATITUDE’ and ‘LONGTITUDE’ column since they are the same. They are in decimal degree format, hence we will assume the data is wgs84.

We will next check if there are any NA value in ‘LATITUDE’ and ‘LONGTITUDE’ using is.na function of base R package.

Show code 
any(is.na(tourism$LATITUDE))

[1] TRUE
Show code 
any(is.na(tourism$LATITUDE))

[1] TRUE

There are NA values and we will need to drop them before creating a simple feature data frame.

The following code chunk drop rows of data as long as the LATITUDE is NA.

Show code 
tourism <- tourism %>%
            drop_na("LATITUDE")

4.2.2.3 Create simple feature data frame from aspatial data

The following code 1. convert tourism data into simple feature data frame using st_as_sf feature of sf package. 2. ‘LONGTITUDE’ and ‘LATITUDE’ column are set as the coords arguments. 3. CRS were set as 4326 since we assumed that the data is wgs84 4. st_transform function of sf package then helps to transform the newly created simple feature data frame into svy21 5. tm_shape and tm_dots function of tmap plots the transformed tourism_sf

Show code 
tourism_sf <- st_as_sf(tourism,
                    coords = c("LONGTITUDE", 
                               "LATITUDE"),
                    crs=4326) %>%
  st_transform(crs = 3414)

tm_shape(tourism_sf)+
  tm_dots()

4.2.3 Airbnb0619

4.2.3.1 Importing

The following code uses read_csv function of readr package to import Airbnb listings30062019.csv aspatial data sets.

Show code 
airbnb0619 <- read_csv("data/aspatial/listings30062019.csv")

4.2.3.2 Examine the file

The following code uses list function of base R to examine the imported file

Show code 
list(airbnb0619)

[[1]]
# A tibble: 8,293 x 16
       id name        host_id host_name neighbourhood_g~ neighbourhood
    <dbl> <chr>         <dbl> <chr>     <chr>            <chr>        
 1  49091 "COZICOMFO~  266763 Francesca North Region     Woodlands    
 2  50646 "Pleasant ~  227796 Sujatha   Central Region   Bukit Timah  
 3  56334 "COZICOMFO~  266763 Francesca North Region     Woodlands    
 4  71609 "Ensuite R~  367042 Belinda   East Region      Tampines     
 5  71896 "B&B  Room~  367042 Belinda   East Region      Tampines     
 6  71903 "Room 2-ne~  367042 Belinda   East Region      Tampines     
 7  71907 "3rd level~  367042 Belinda   East Region      Tampines     
 8 151547 "En-suite ~  729617 Kanchuya  Central Region   Orchard      
 9 241503 "Long stay~ 1017645 Bianca    East Region      Bedok        
10 241508 "Long stay~ 1017645 Bianca    East Region      Bedok        
# ... with 8,283 more rows, and 10 more variables: latitude <dbl>,
#   longitude <dbl>, room_type <chr>, price <dbl>,
#   minimum_nights <dbl>, number_of_reviews <dbl>,
#   last_review <date>, reviews_per_month <dbl>,
#   calculated_host_listings_count <dbl>, availability_365 <dbl>

The airbnb0619 data contains 8293 rows and 16 columns. There are ‘latitude’ and ‘longitude’ column for coordinates. They are in decimal degree format, hence we will assume the data is wgs84.

We will next check if there are any NA value in ‘latitude’ and ‘longitude’ using is.na function of base R package.

Show code 
any(is.na(airbnb0619$latitude))

[1] FALSE
Show code 
any(is.na(airbnb0619$longitude))

[1] FALSE

Since there are no NA values, we can proceed to create simple features data frame for airbnb0619

4.2.3.3 Create simple feature data frame from aspatial data

The following code 1. convert airbnb0619 data into simple feature data frame using st_as_sf feature of sf package. 2. ‘longitude’ and ‘latitude’ column are set as the coords arguments. 3. CRS were set as 4326 since we assumed that the data is wgs84 4. st_transform function of sf package then helps to transform the newly created simple feature data frame into svy21 5. tm_shape and tm_dots function of tmap plots the transformed airbnb0619_sf

Show code 
airbnb0619_sf <- st_as_sf(airbnb0619,
                    coords = c("longitude", 
                               "latitude"),
                    crs=4326) %>%
  st_transform(crs = 3414)

tm_shape(airbnb0619_sf)+
  tm_dots()

4.2.4 Airbnb0721

4.2.4.1 Importing

The following code uses read_csv function of readr package to import Airbnb listings19072021.csv aspatial data sets.

Show code 
airbnb0721 <- read_csv("data/aspatial/listings19072021.csv")

4.2.4.2 Examine the file

The following code uses list function of base R to examine the imported file

Show code 
list(airbnb0721)

[[1]]
# A tibble: 4,252 x 16
       id name        host_id host_name neighbourhood_g~ neighbourhood
    <dbl> <chr>         <dbl> <chr>     <chr>            <chr>        
 1  50646 Pleasant R~  227796 Sujatha   Central Region   Bukit Timah  
 2  71609 Ensuite Ro~  367042 Belinda   East Region      Tampines     
 3  71896 B&B  Room ~  367042 Belinda   East Region      Tampines     
 4  71903 Room 2-nea~  367042 Belinda   East Region      Tampines     
 5 275343 Convenient~ 1439258 Joyce     Central Region   Bukit Merah  
 6 275344 15 mins to~ 1439258 Joyce     Central Region   Bukit Merah  
 7 294281 5 mins wal~ 1521514 Elizabeth Central Region   Newton       
 8 301247 Nice room ~ 1552002 Rahul     Central Region   Geylang      
 9 324945 20 Mins to~ 1439258 Joyce     Central Region   Bukit Merah  
10 330089 Accomo@ RE~ 1439258 Joyce     Central Region   Bukit Merah  
# ... with 4,242 more rows, and 10 more variables: latitude <dbl>,
#   longitude <dbl>, room_type <chr>, price <dbl>,
#   minimum_nights <dbl>, number_of_reviews <dbl>,
#   last_review <date>, reviews_per_month <dbl>,
#   calculated_host_listings_count <dbl>, availability_365 <dbl>

The airbnb0721 data contains 4252 rows and 16 columns, lesser than airbnb0619. There are ‘latitude’ and ‘longitude’ column for coordinates. They are in decimal degree format, hence we will assume the data is wgs84.

We will next check if there are any NA value in ‘latitude’ and ‘longitude’ using is.na function of base R package.

Show code 
any(is.na(airbnb0721$latitude))

[1] FALSE
Show code 
any(is.na(airbnb0721$longitude))

[1] FALSE

We can confirm that there are no latitude and longitude with NA value.

4.2.4.3 Create simple feature data frame from aspatial data

The following code 1. convert airbnb0721 data into simple feature data frame using st_as_sf feature of sf package. 2. ‘longitude’ and ‘latitude’ column are set as the coords arguments. 3. CRS were set as 4326 since we assumed that the data is wgs84 4. st_transform function of sf package then helps to transform the newly created simple feature data frame into svy21 5. tm_shape and tm_dots function of tmap plots the transformed airbnb0721_sf

Show code 
airbnb0721_sf <- st_as_sf(airbnb0721,
                    coords = c("longitude", 
                               "latitude"),
                    crs=4326) %>%
  st_transform(crs = 3414)

tm_shape(airbnb0721_sf)+
  tm_dots()

4.3 Geospatial Data wrangling

After importing geospatial and aspatial data, we will convert the simple feature data frame to sp’s Spatial* class.

4.3.1 Converting sf data frames to sp’s Spatial* class

The following code chunk converts all the geospatial data from simple feature data frame to sp’s Spatial* class using as_Spatial functions of sf package.

Show code 
airbnb0619 <- as_Spatial(airbnb0619_sf)
airbnb0721 <- as_Spatial(airbnb0721_sf)
hotels <- as_Spatial(hotels_sf)
tourism <- as_Spatial(tourism_sf)
mrtlrt <- as_Spatial(mrtlrt_sf)
sg <- as_Spatial(sg_sf)
mpsz <- as_Spatial(mpsz_sf)

Check each Spatial* class

Show code 
airbnb0619

class       : SpatialPointsDataFrame 
features    : 8293 
extent      : 7215.566, 44098.31, 25166.35, 49226.35  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 14
names       :       id,                                              name,   host_id,                host_name, neighbourhood_group, neighbourhood,       room_type, price, minimum_nights, number_of_reviews, last_review, reviews_per_month, calculated_host_listings_count, availability_365 
min values  :    49091,                                                 -,     23666, (Email hidden by Airbnb),      Central Region,    Ang Mo Kio, Entire home/apt,     0,              1,                 0,       15656,              0.01,                              1,                0 
max values  : 36053005, ZR2- NEW! Sunny & Modern Apt 4 mins to Orchard Rd, 271165196,                   Zuzana,         West Region,        Yishun,     Shared room, 13999,           1000,               308,       18072,             12.09,                            277,              365
Show code 
airbnb0721

class       : SpatialPointsDataFrame 
features    : 4252 
extent      : 7406.989, 43337.89, 25330, 48391.55  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 14
names       :       id,                                               name,   host_id,        host_name, neighbourhood_group, neighbourhood,       room_type, price, minimum_nights, number_of_reviews, last_review, reviews_per_month, calculated_host_listings_count, availability_365 
min values  :    50646, !! CozyRoom@City Center,Little India,FarrerParkMRT,     23666, <U+4F73><U+4F73>,      Central Region,    Ang Mo Kio, Entire home/apt,    13,              1,                 0,       15456,              0.01,                              1,                0 
max values  : 51122424,                 Zimmer nah der MRT (Downtown Line), 411376818,             Zuzu,         West Region,        Yishun,     Shared room, 13400,           1000,               370,       18826,                77,                            165,              365
Show code 
hotels

class       : SpatialPointsDataFrame 
features    : 422 
extent      : 5939.241, 45334.18, 25379.44, 44562.4  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 7
names       :                              NAME, ADDRESSPOSTALCODE,                                 ADDRESSSTREETNAME,         HYPERLINK, TOTALROOMS,     KEEPERNAME, ICON_NAME 
min values  :                      30 BENCOOLEN,             18956,                                 1 Bayfront Avenue, 96ytlim@gmail.com,          4,  Adel Aramouni, hotel.gif 
max values  : YotelAir Singapore Changi Airport,            819666, 99 IRRAWADDY ROAD, # 22-00 ROYAL SQUARE AT NOVENA, zubair@dam.com.sg,       2561, Zhang YuanQing, hotel.gif
Show code 
tourism

class       : SpatialPointsDataFrame 
features    : 106 
extent      : 11380.23, 43659.54, 22869.34, 47596.73  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 15
names       :                                NAME,                                                                                                                                                  DESCRIPTION,                  ADDRESSSTREETNAME,                            HYPERLINK,                                                                                                                             PHOTOURL,                                                                                          URL_PATH,                                                                                                          IMAGE_ALT_TEXT,                 PHOTOCREDITS,   LASTMODIFIED,                                                                                             META_DESCRIPTION,                                                                                                                                                                           OPENING_HOURS,              Lat,              Lng,        ICON_NAME, ADDRESSPOSTALCODE 
min values  : Adventure Cove Waterpark™ Singapore, A feat of engineering, an architectural statement and a sheer aesthetic triumph, Marina Bay Sands<sup>®</sup> has upped the ante for buildings in Singapore.,                       1 Beach Road,                   http://acm.org.sg/,                                  /content/dam/desktop/global/see-do-singapore/architecture/hajjah-fatimah-mosque-carousel01-rect.jpg, www.yoursingapore.com/en/see-do-singapore/architecture/historical/capitol-building-singapore.html,               Adults and kids of all ages who are not even science buffs will have fun at the Singapore Science Centre., ©Darren Soh/National Gallery, 1427691447.648, A tranquil patch of imperial China in the west of Singapore is pleasant respite from the bustle of the city.,                                                                                                                                                 50th storey Skybridge, Daily, 9am –10pm, 1.22309649975255, 103.683980000372, tourist_spot.gif,                 0 
max values  :          Victoria Theatre Singapore,                                                    With so many attractions packed into this 15-km stretch of beaches, you’ll never run out of things to do., Seng Poh Road and Tiong Bahru Road, https://www.pub.gov.sg/marinabarrage, www.yoursingapore.com/content/dam/desktop/global/see-do-singapore/recreation-leisure/universal-studios-singapore-carousel01-rect.jpg,      www.yoursingapore.com/en/see-do-singapore/recreation-leisure/viewpoints/singapore-flyer.html, Whether you prefer water sports, rollerblading or cycling, find a myriad of things to do at East Coast Park, Singapore.,  Wildlife Reserves Singapore, 1446544541.364,                                     With the Henderson Waves bridge, form meets function to stunning effect., Visits are by appointment only.Visitors must sign up in advance for heritage tours which fall on:Monday, 2pm – 3pm,Tuesday, 6.30pm – 7.30pm,Thursday, 10am – 11am,Saturday, 11am – 12pm, 1.44672000006097, 103.974030000359, tourist_spot.gif,                 0
Show code 
mrtlrt

class       : SpatialPointsDataFrame 
features    : 185 
extent      : 6138.311, 45254.86, 27555.06, 47854.2  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 3
names       : OBJECTID,              STN_NAME, STN_NO 
min values  :        0, ADMIRALTY MRT STATION,    BP1 
max values  :      191,    YISHUN MRT STATION,    TE3
Show code 
sg

class       : SpatialPolygonsDataFrame 
features    : 60 
extent      : 2663.926, 56047.79, 16357.98, 50244.03  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 4
names       : GDO_GID, MSLINK, MAPID,              COSTAL_NAM 
min values  :       1,      1,     0,             ISLAND LINK 
max values  :      60,     67,     0, SINGAPORE - MAIN ISLAND
Show code 
mpsz

class       : SpatialPolygonsDataFrame 
features    : 323 
extent      : 2667.538, 56396.44, 15748.72, 50256.33  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 15
names       : OBJECTID, SUBZONE_NO, SUBZONE_N, SUBZONE_C, CA_IND, PLN_AREA_N, PLN_AREA_C,       REGION_N, REGION_C,          INC_CRC, FMEL_UPD_D,     X_ADDR,     Y_ADDR,    SHAPE_Leng,    SHAPE_Area 
min values  :        1,          1, ADMIRALTY,    AMSZ01,      N, ANG MO KIO,         AM, CENTRAL REGION,       CR, 00F5E30B5C9B7AD8,      16409,  5092.8949,  19579.069, 871.554887798, 39437.9352703 
max values  :      323,         17,    YUNNAN,    YSSZ09,      Y,     YISHUN,         YS,    WEST REGION,       WR, FFCCF172717C2EAF,      16409, 50424.7923, 49552.7904, 68083.9364708,  69748298.792

We can see that all the geospatial data are converted to their respective Spatial* Classes.

4.3.2 Converting the Spatial* class into generic sp format

Next, we will convert Spatial* classes into Spatial object so that we can convert it into ppp object.

The following code chunk converts Spatial* classes into generic sp format using as function which sets the classes to their respective sp format.

Show code 
airbnb0619_sp <- as(airbnb0619, "SpatialPoints")
airbnb0721_sp <- as(airbnb0721, "SpatialPoints")
hotels_sp <- as(hotels, "SpatialPoints")
tourism_sp <- as(tourism, "SpatialPoints")
mrtlrt_sp <- as(mrtlrt, "SpatialPoints")
sg_sp <- as(sg, "SpatialPolygons")
mpsz_sp <- as(mpsz, "SpatialPolygons")

4.3.3 Converting the generic sp format into spatstat’s ppp format

We will next convert the spatial data into spatstat’s ppp format.

The following code chunk converts spatial data into spatstat’s ppp format using as function.

Show code 
airbnb0619_ppp <- as(airbnb0619_sp, "ppp")
airbnb0721_ppp <- as(airbnb0721_sp, "ppp")
hotels_ppp <- as(hotels_sp, "ppp")
tourism_ppp <- as(tourism_sp, "ppp")
mrtlrt_ppp <- as(mrtlrt_sp, "ppp")

4.3.4 Handling duplicates in ppp object

4.3.4.1 Check for dupliciates duplicates in each ppp object

The following code chunk use any and duplicated function of base R package to check for duplicates, and sum up the count of duplicated points using sum function of base R package and multiplicity function of spatstat package.

Show code 
any(duplicated(airbnb0619_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0619_ppp) > 1)

[1] 6

From the result above, there are 6 duplicates in airbnb0619_ppp object.

Show code 
any(duplicated(airbnb0721_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0721_ppp) > 1)

[1] 224

From the result above, there are 224 duplicates in airbnb0721_ppp object.

Show code 
any(duplicated(hotels_ppp))

[1] TRUE
Show code 
sum(multiplicity(hotels_ppp) > 1)

[1] 10

From the result above, there are 10 duplicates in hotels_ppp object.

Show code 
any(duplicated(tourism_ppp))

[1] TRUE
Show code 
sum(multiplicity(tourism_ppp) > 1)

[1] 7

From the result above, there are 7 duplicates in tourism_ppp object.

Show code 
any(duplicated(mrtlrt_ppp))

[1] FALSE
Show code 
sum(multiplicity(mrtlrt_ppp) > 1)

[1] 0

From the result above, there are no duplicate in mrtlrt_ppp object.

4.3.4.2 Handling duplicates

We will use jittering to handle duplicates so that we will not lose useful points. Jittering adds a small perturbation to the duplicated points so that they do not occupy the exact same space.

The following code chunk uses rjitter function of spatstat to jitter the ppp object with duplicates.

Show code 
airbnb0619_ppp_jit <- rjitter(airbnb0619_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

airbnb0721_ppp_jit <- rjitter(airbnb0721_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

hotels_ppp_jit <- rjitter(hotels_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

tourism_ppp_jit <- rjitter(tourism_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

4.3.4.3 Check if duplicates still exist

Show code 
any(duplicated(airbnb0619_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0619_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(airbnb0721_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0721_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(hotels_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(hotels_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(tourism_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(tourism_ppp_jit) > 1)

[1] 0

We can confirm that there are no duplicated points after jittering.

4.3.5 Create owin objects

We will next create owin object to represent the Singapore boundary.

The following code chunk converts sg SpatialPolygon objects into owin object of spatstat using as function.

Show code 
sg_owin <- as(sg_sp, "owin")

The following code chunk plots sg_owin using plot function of base R package.

Show code 
plot(sg_owin)

4.3.6 Combining point events object and owin object

The following code chunk combines the point objects of airbnb0619, airbnb0721, hotels, tourism and mrtlrt with the owin object into one ppp object class.

Show code 
airbnb0619SG_ppp = airbnb0619_ppp_jit[sg_owin]
airbnb0721SG_ppp = airbnb0721_ppp_jit[sg_owin]
hotelsSG_ppp = hotels_ppp_jit[sg_owin]
tourismSG_ppp = tourism_ppp_jit[sg_owin]
mrtlrtSG_ppp = mrtlrt_ppp[sg_owin]

The following code chunk plots the ppp object class using plot function of base R package.

Show code 
plot(airbnb0619SG_ppp)

Show code 
plot(airbnb0721SG_ppp)

Show code 
plot(hotelsSG_ppp)

Show code 
plot(tourismSG_ppp)

Show code 
plot(mrtlrtSG_ppp)

5 Section A: Airbnb Distribution in 2019

5.1 Exploratory Spatial Data Analysis

5.1.1 Kernel density maps

5.1.1.1 Airbnb0619 listings

5.1.1.1.1 Compute kernel density estimation with automatic bandwidth selection method
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
Show code 
kde_airbnb0619SG_bw <- density(airbnb0619SG_ppp,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

The following code chunk plots the kde_airbnb0619SG_bw using plot function of base R

Show code 
plot(kde_airbnb0619SG_bw)

The density values of the output range is too small because the default unit of measurement of svy21 is in meter. As a result, the density values computed is in “number of points per square meter”.

5.1.1.1.2 Rescalling KDE values

The following code chunk rescale the KDE value of airbnb0619SG_ppp using rescale function of spatstat from meter to kilometer.

Show code 
airbnb0619SG_ppp.km <- rescale(airbnb0619SG_ppp, 1000, "km")
5.1.1.1.3 Different automatic bandwidth methods

The following code chunk calculate the bandwidth using different methods

Show code 
bw.CvL(airbnb0619SG_ppp.km)

   sigma 
6.238907
Show code 
bw.scott(airbnb0619SG_ppp.km)

 sigma.x  sigma.y 
1.074846 0.742538
Show code 
bw.ppl(airbnb0619SG_ppp.km)

    sigma 
0.2955613
Show code 
bw.diggle(airbnb0619SG_ppp.km)

    sigma 
0.0419688

From the result above, since bw.diggle has the lowest sigma, we will be using it for the other ppp object in order to detect close cluster.

5.1.1.1.4 KDE Map and Point Map
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable in km.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
  5. plot function of base R package and par function of graphics package is used to plot kde map and ppp point map side by side
Show code 
kde_airbnb0619SG_bw.km <- density(airbnb0619SG_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")


par(mfrow=c(1,2))
plot(kde_airbnb0619SG_bw.km)
plot(airbnb0619SG_ppp)

Analysis: From the kde map and point map above, we can see that the south-east region near the central area have higher density of airbnb in 2019 as shown by the pink spots on kde map and the black patches on the point map. The kde map better visualised the density as compared to the point map because it uses different colour tone to represent the density, for example, purple means no/low density, pink represents density of 1000 and yellow means density of around 1500. The point map on the other hand is not a very good visualization because we cannot determine if an area is denser with one colour tone (black, grey and white).

5.1.1.1.5 Convert KDE output into grid object

The following code chunk converts KDE output into grid object using as.SpatialGridDataFrame.im function of maptools, and plot the grid object using spplot function of raster package.

Show code 
gridded_kde_airbnb0619SG_bw <- as.SpatialGridDataFrame.im(kde_airbnb0619SG_bw.km)
spplot(gridded_kde_airbnb0619SG_bw)

Analysis: The above density map shows a clearer visualization of the density of airbnb in 2019 using grid. We can see that the airbnb are scattered all around as seen on the lighter blue grid on the map. The south-east area near the central have high density as seen on the pink to yellow grid on the map.

5.1.1.1.6 Convert grid object into raster

The following code chunk converts grid object into raster using raster function of raster package

Show code 
kde_airbnb0619SG_bw_raster <- raster(gridded_kde_airbnb0619SG_bw)
kde_airbnb0619SG_bw_raster

class      : RasterLayer 
dimensions : 128, 128, 16384  (nrow, ncol, ncell)
resolution : 0.4170614, 0.2647348  (x, y)
extent     : 2.663926, 56.04779, 16.35798, 50.24403  (xmin, xmax, ymin, ymax)
crs        : NA 
source     : memory
names      : v 
values     : -2.270034e-13, 1784.248  (min, max)

From the result above, the crs is NA.

5.1.1.1.7 Assign project system and plot raster using tmap

The following code chunk assign CRS to kde_airbnb0619SG_bw_raster and plot the raster using tmap package

Show code 
projection(kde_airbnb0619SG_bw_raster) <- CRS("+init=EPSG:3414")
kde_airbnb0619SG_bw_raster

class      : RasterLayer 
dimensions : 128, 128, 16384  (nrow, ncol, ncell)
resolution : 0.4170614, 0.2647348  (x, y)
extent     : 2.663926, 56.04779, 16.35798, 50.24403  (xmin, xmax, ymin, ymax)
crs        : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +units=m +no_defs 
source     : memory
names      : v 
values     : -2.270034e-13, 1784.248  (min, max)
Show code 
tmap_mode("plot")
tm_shape(kde_airbnb0619SG_bw_raster) + 
  tm_raster("v") +
  tm_layout(legend.position = c("right", "bottom"), frame = FALSE)

Analysis: From the raster map, we can confirm that the south-east area near the central have the highest density of airbnb as shown by the green spot on the map.

5.1.1.2 Hotels

5.1.1.2.1 Rescale KDE values

The following code chunk rescale the KDE value of hotelsSG_ppp using rescale function of spatstat from meter to kilometer.

Show code 
hotelsSG_ppp.km <- rescale(hotelsSG_ppp, 1000, "km")
5.1.1.2.2 KDE Map and Point Map
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable in km.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
  5. plot function of base R package and par function of graphics package is used to plot kde map and ppp point map side by side
Show code 
kde_hotelsSG_bw.km <- density(hotelsSG_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

par(mfrow=c(1,2))
plot(kde_hotelsSG_bw.km, main = "KDE Map")
plot(hotelsSG_ppp, main="Point Map")

Analysis: From the kde map and point map above, we can see that the south-east region near the central area have higher density of hotels as shown by the pink orange spots on kde map and the black patch on the point map.

5.1.1.2.3 Convert KDE output into grid object

The following code chunk converts KDE output into grid object using as.SpatialGridDataFrame.im function of maptools, and plot the grid object using spplot function of raster package.

Show code 
gridded_kde_hotelsSG_bw <- as.SpatialGridDataFrame.im(kde_hotelsSG_bw.km)
spplot(gridded_kde_hotelsSG_bw)

Analysis: The above density map shows a clearer visualization of the density of hotels using grid. We can see that south-east area near the central have higher density as seen on the pink grid on the map, and there are also many light blue grid at the south-east region near central area. There are some hotels (individual blue grid) scatter near east and west area as well.

5.1.1.2.4 Convert grid object into raster, assign project system and plot raster

The following code chunk converts grid object into raster using raster function of raster package, and plot raster with tmap.

Show code 
kde_hotelsSG_bw_raster <- raster(gridded_kde_hotelsSG_bw)
projection(kde_hotelsSG_bw_raster) <- CRS("+init=EPSG:3414")

tmap_mode("plot")
tm_shape(kde_hotelsSG_bw_raster) + 
  tm_raster("v") +
  tm_layout(legend.position = c("right", "bottom"), frame = FALSE)

Analysis: From the raster map, we can confirm that the south-east area near the central have the highest density of hotels as shown by the green spot on the map.

5.1.1.3 MRTLRT

5.1.1.3.1 Rescale KDE values

The following code chunk rescale the KDE value of mrtlrtSG_ppp using rescale function of spatstat from meter to kilometer.

Show code 
mrtlrtSG_ppp.km <- rescale(mrtlrtSG_ppp, 1000, "km")
5.1.1.3.2 KDE Map and Point Map
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable in km.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
  5. plot function of base R package and par function of graphics package is used to plot kde map and ppp point map side by side
Show code 
kde_mrtlrtSG_bw.km <- density(mrtlrtSG_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

par(mfrow=c(1,2))
plot(kde_mrtlrtSG_bw.km, main = "KDE Map")
plot(mrtlrtSG_ppp, main="Point Map")

Analysis: From the kde map and point map above, we can see that the mrt and lrt are available at almost every region. Area with higher density of mrt and lrt are North-East, North-West and Central South region as shown by the pink orange patch on the kde map and the black patches on the point map.

5.1.1.3.3 Convert KDE output into grid object

The following code chunk converts KDE output into grid object using as.SpatialGridDataFrame.im function of maptools, and plot the grid object using spplot function of raster package.

Show code 
gridded_kde_mrtlrtSG_bw <- as.SpatialGridDataFrame.im(kde_mrtlrtSG_bw.km)
spplot(gridded_kde_mrtlrtSG_bw)

Analysis: The above density map shows a clearer visualization of the density of Mrt and Lrt using grid. We can see that the Mrt and Lrt are available almost every region in Singapore as represented by the light blue grid on the map. The north-east, north-west and central south area have higher density of Mrt and Lrt as seen by the pink to yellow grid cluster on the map.

5.1.1.3.4 Convert grid object into raster, assign project system and plot raster

The following code chunk converts grid object into raster using raster function of raster package, and plot raster using tmap.

Show code 
kde_mrtlrtSG_bw_raster <- raster(gridded_kde_mrtlrtSG_bw)
projection(kde_mrtlrtSG_bw_raster) <- CRS("+init=EPSG:3414")

tmap_mode("plot")
tm_shape(kde_mrtlrtSG_bw_raster) + 
  tm_raster("v") +
  tm_layout(legend.position = c("right", "bottom"), frame = FALSE)

Analysis: From the raster map above, we can confirm that north-east, north-west and central south area have higher density of Mrt and Lrt as shown by the green spots on the map.

5.1.1.4 Tourist Attractions

5.1.1.4.1 Rescale KDE values

The following code chunk rescale the KDE value of mrtlrtSG_ppp using rescale function of spatstat from meter to kilometer.

Show code 
tourismSG_ppp.km <- rescale(tourismSG_ppp, 1000, "km")
5.1.1.4.2 KDE Map and Point Map
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable in km.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
  5. plot function of base R package and par function of graphics package is used to plot kde map and ppp point map side by side
Show code 
kde_tourismSG_bw.km <- density(tourismSG_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

par(mfrow=c(1,2))
plot(kde_tourismSG_bw.km, main = "KDE Map")
plot(tourismSG_ppp, main="Point Map")

Analysis: From the kde map and point map above, we can see that the south-east region near the central area have higher density of tourist attraction as shown on the pink spots in kde map and the black patches in the point map. In addition, there are some tourist attractions in north, west and east as shown by the individual point on the point map.

5.1.1.4.3 Convert KDE output into grid object

The following code chunk converts KDE output into grid object using as.SpatialGridDataFrame.im function of maptools, and plot the grid object using spplot function of raster package.

Show code 
gridded_kde_tourismSG_bw <- as.SpatialGridDataFrame.im(kde_tourismSG_bw.km)
spplot(gridded_kde_tourismSG_bw)

Analysis: The above density map shows a clearer visualization of the density of tourist attraction in Singapore using grid. We can see that there are a few attractions in the north, west and east as seen by the light blue grid on the map. The south central area have high density of tourist attractions as seen by the pink yellow grid on the map.

5.1.1.4.4 Convert grid object into raster, assign project system and plot raster

The following code chunk converts grid object into raster using raster function of raster package, and plot raster using tmap.

Show code 
kde_tourismSG_bw_raster <- raster(gridded_kde_tourismSG_bw)
projection(kde_tourismSG_bw_raster) <- CRS("+init=EPSG:3414")

tmap_mode("plot")
tm_shape(kde_tourismSG_bw_raster) + 
  tm_raster("v") +
  tm_layout(legend.position = c("right", "bottom"), frame = FALSE)

Analysis: From the raster map above, we can confirm the the south-east area near central region have highest density of tourist attraction as shown by the green spots on the map.

5.1.1.5 Comparison between Airbnb0619, Hotels, MRTLRT and Tourist Attractions

5.1.1.5.1 KDE Map Comparison

The following code chunk plots all the KDE map of Airbnb 2019, Hotel, MRTLRT and Tourist Attractions side by side using par functions of graphics package

Show code 
par(mfrow=c(2,2))
plot(kde_airbnb0619SG_bw.km, main = "Airbnb 2019")
plot(kde_hotelsSG_bw.km, main = "Hotel")
plot(kde_mrtlrtSG_bw.km, main = "Mrt")
plot(kde_tourismSG_bw.km, main = "Tourist Attraction")

From the above, we can see that the common area with high density (in red and yellow) of airbnb, hotel, mrt and tourist attraction are concentrated around the south east area near the central region.

5.1.1.5.2 Central Region

We will look into the central region

5.1.1.5.2.1 Exploration with OpenStreetMap

The following code chunk first extract airbnb in the central region from airbnb0619_sf region.

tmap_mode is set to “view” for interaction mode and tm_basemap is set to “OpenStreetMap”

centralairbnb0619 are plotted using tm_shape and tm_dots of tmap packages.

tm_basemap function sets OpenStreetMap as the basemap.

Show code 
centralairbnb0619 <- airbnb0619_sf[airbnb0619_sf$neighbourhood_group == "Central Region",]

tmap_mode("view")

tm_shape(centralairbnb0619) + 
  tm_dots() +
tm_basemap(leaflet::providers$OpenStreetMap)

From the map above, we can see that there are many Airbnbs located in Kallang, River Valley, Novena and Rochor.

Show code 
tmap_mode("plot")
5.1.1.5.2.2 Comparing Spatial Point Patterns using KDE

Extract Study Area

The following code chunk extracts the target areas (Kallang, River Valley, Novena, Rochor).

Show code 
kl = mpsz[mpsz@data$PLN_AREA_N == "KALLANG",]
rv = mpsz[mpsz@data$PLN_AREA_N == "RIVER VALLEY",]
nv = mpsz[mpsz@data$PLN_AREA_N == "NOVENA",]
rc = mpsz[mpsz@data$PLN_AREA_N == "ROCHOR",]

Plot target areas

The following code chunk plots the target areas side by side in 2x2 using par function of graphics package and plot functions of base R package.

Show code 
par(mfrow=c(2,2))
plot(kl, main = "Kallang")
plot(rv, main = "River Valley")
plot(nv, main = "Novena")
plot(rc, main = "Rochor")

Convert Spatial Polygons Data Fraeme into generic SP format

The following code chunk converts the four SpatialPolygonsDataFrame layers into generic spatialpolygons layers.

Show code 
kl_sp = as(kl, "SpatialPolygons")
rv_sp = as(rv, "SpatialPolygons")
nv_sp = as(nv, "SpatialPolygons")
rc_sp = as(rc, "SpatialPolygons")

Create owin object for each target area

The following code chunk converst the SpatialPolygons objects into owin objects required by spatstat.

Show code 
kl_owin = as(kl_sp, "owin")
rv_owin = as(rv_sp, "owin")
nv_owin = as(nv_sp, "owin")
rc_owin = as(rc_sp, "owin")

Combine airnbnb0619, hotels, mrt, tourist attractions points with owin of each target area

The following code chunk extracts airnbnb0619, hotels, mrt and tourist attractions within each specific region.

Show code 
airbnb0619_kl_ppp = airbnb0619_ppp_jit[kl_owin]
airbnb0619_rv_ppp = airbnb0619_ppp_jit[rv_owin]
airbnb0619_nv_ppp = airbnb0619_ppp_jit[nv_owin]
airbnb0619_rc_ppp = airbnb0619_ppp_jit[rc_owin]

hotels_kl_ppp = hotels_ppp_jit[kl_owin]
hotels_rv_ppp = hotels_ppp_jit[rv_owin]
hotels_nv_ppp = hotels_ppp_jit[nv_owin]
hotels_rc_ppp = hotels_ppp_jit[rc_owin]

mrtlrt_kl_ppp = mrtlrt_ppp[kl_owin]
mrtlrt_rv_ppp = mrtlrt_ppp[rv_owin]
mrtlrt_nv_ppp = mrtlrt_ppp[nv_owin]
mrtlrt_rc_ppp = mrtlrt_ppp[rc_owin]

tourism_kl_ppp = tourism_ppp_jit[kl_owin]
tourism_rv_ppp = tourism_ppp_jit[rv_owin]
tourism_nv_ppp = tourism_ppp_jit[nv_owin]
tourism_rc_ppp = tourism_ppp_jit[rc_owin]

Rescale from m to km

The following code chunk transform the unit from m to km using rescale function.

Show code 
airbnb0619_kl_ppp.km = rescale(airbnb0619_kl_ppp, 1000, "km")
airbnb0619_rv_ppp.km = rescale(airbnb0619_rv_ppp, 1000, "km")
airbnb0619_nv_ppp.km = rescale(airbnb0619_nv_ppp, 1000, "km")
airbnb0619_rc_ppp.km = rescale(airbnb0619_rc_ppp, 1000, "km")

hotels_kl_ppp.km = rescale(hotels_kl_ppp, 1000, "km")
hotels_rv_ppp.km = rescale(hotels_rv_ppp, 1000, "km")
hotels_nv_ppp.km = rescale(hotels_nv_ppp, 1000, "km")
hotels_rc_ppp.km = rescale(hotels_rc_ppp, 1000, "km")

mrtlrt_kl_ppp.km = rescale(mrtlrt_kl_ppp, 1000, "km")
mrtlrt_rv_ppp.km = rescale(mrtlrt_rv_ppp, 1000, "km")
mrtlrt_nv_ppp.km = rescale(mrtlrt_nv_ppp, 1000, "km")
mrtlrt_rc_ppp.km = rescale(mrtlrt_rc_ppp, 1000, "km")

tourism_kl_ppp.km = rescale(tourism_kl_ppp, 1000, "km")
tourism_rv_ppp.km = rescale(tourism_rv_ppp, 1000, "km")
tourism_nv_ppp.km = rescale(tourism_nv_ppp, 1000, "km")
tourism_rc_ppp.km = rescale(tourism_rc_ppp, 1000, "km")

Airbnb comparison by neighbourhood

The following code chunk plots airbnb within each neigbourhood using plot and density functions with sigma set to bw.diggle and kernel set to gaussian.

Show code 
par(mfrow=c(2,2))
plot(density(airbnb0619_kl_ppp.km, 
             sigma=bw.diggle, 
             edge=TRUE, 
             kernel="gaussian"),
     main="Kallang")
plot(density(airbnb0619_rv_ppp.km, 
             sigma=bw.diggle, 
             edge=TRUE, 
             kernel="gaussian"),
     main="River Valley")
plot(density(airbnb0619_nv_ppp.km, 
             sigma=bw.diggle, 
             edge=TRUE, 
             kernel="gaussian"),
     main="Novena")
plot(density(airbnb0619_rc_ppp.km, 
             sigma=bw.diggle, 
             edge=TRUE, 
             kernel="gaussian"),
     main="Rochor")

Analysis: From the comparison, we noticed that the scale range for each region are different. For example, Kallang have a larger scale range from 0 - 2500 as compared to Rochor with a smaller scale range of 100 - 700. In Kallang, Airbnb are denser around the north-west area with high density of around 1500 to 2500. In River Valley, Airbnb are denser at the south-east area with high density of around 600 to 1400. In Novena, the east area have higher density of Airbnb of around 600 to 1200. In Rochor, the outer areas of the region have high density of Airbnb at about 400 to 700.

Comparison within Kallang

The following code chunk plots airbnb, hotel, mrt and tourist attraction in Kallang.

Show code 
par(mfrow=c(2,2))
plot(airbnb0619_kl_ppp.km, main = "Airbnb")
plot(hotels_kl_ppp.km, main = "Hotel")
plot(mrtlrt_kl_ppp.km, main = "Mrt")
plot(tourism_kl_ppp.km, main = "Tourist Attraction")

Analysis: Even though there are some hotels, there are still many airbnb in Kallang. There are a few MRT stations and tourist attrations in Kallang, which may be the reason to the many airbnb located in this area.

Comparison within River Valley

The following code chunk plots airbnb, hotel, mrt and tourist attraction in River Valley.

Show code 
par(mfrow=c(2,2))
plot(airbnb0619_rv_ppp.km, main = "Airbnb")
plot(hotels_rv_ppp.km, main = "Hotel")
plot(mrtlrt_rv_ppp.km, main = "Mrt")
plot(tourism_rv_ppp.km, main = "Tourist Attraction")

Analysis: Even though there are no tourist attraction and Mrt in River Valley, there are still many airbnb. There are only one hotel and it may be the reason for the large number of airbnb in the area since there are lesser competition.

Comparison within Novena

The following code chunk plots airbnb, hotel, mrt and tourist attraction in Novena.

Show code 
par(mfrow=c(2,2))
plot(airbnb0619_nv_ppp.km, main = "Airbnb")
plot(hotels_nv_ppp.km, main = "Hotel")
plot(mrtlrt_nv_ppp.km, main = "Mrt")
plot(tourism_nv_ppp.km, main = "Tourist Attraction")

Analysis: Even though there are a few hotel in Novena, there are many airbnb. The accessibility of the two MRT and one tourist attraction can be the reason to the high number of airbnb in Novena.

Comparison within Rochor

The following code chunk plots airbnb, hotel, mrt and tourist attraction in Rochor.

Show code 
par(mfrow=c(2,2))
plot(airbnb0619_rc_ppp.km, main = "Airbnb")
plot(hotels_rc_ppp.km, main = "Hotel")
plot(mrtlrt_rc_ppp.km, main = "Mrt")
plot(tourism_rc_ppp.km, main = "Tourist Attraction")

Analysis: In Rochor, there are still many airbnb despite many hotels already available. The accessibility to a few MRTs and multiple tourist attractions may be the reason that many airbnb are located in Rochor.

Overall, we can conclude that there are no direct relationship between airbnb and hotel, mrt and tourist attraction. There are region like River Valley with no tourist attractions and MRT and yet, there are still a high number of airbnb in that area. In addition, regardless of the number of hotel available in a region, we identified that the number of airbnb still remained high in these regions.

5.2 Second-order Spatial Point Patterns Analysis

5.2.1 Nearest Neighbour Analysis

5.2.1.1 Hypothesis

We will perform the Clark-Evans test of aggregation for a spatial point pattern using clarkevans.test() of statspat.

Test Hypothesis:

H0 = The distribution of airbnbs are randomly distributed.

H1 = The distribution of airbnbs are not randomly distributed.

The 95% confident interval will be used.

5.2.1.2 Testing and Analysis

5.2.1.2.1 Testing spatial point patterns using Clark and Evans Test
  1. The following code chunk performs Clark-Evans test using clarkevans.test() of statspat.
  2. The spatial point pattern object is airbnb0619SG_ppp, Singapore airbnb listings in Singapore ppp object.
  3. Correction is set to none
  4. The clipping region is set to sg_owin
  5. The type of alternative for the hypothesis is clustered.
  6. The number of simulations is set to 99.
Show code 
clarkevans.test(airbnb0619SG_ppp,
                correction="none",
                clipregion="sg_owin",
                alternative=c("clustered"),
                nsim=99)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 99 simulations of CSR with fixed n

data:  airbnb0619SG_ppp
R = 0.36024, p-value = 0.01
alternative hypothesis: clustered (R < 1)

Analysis: R = 0.36 is less than 1, which means the distribution of airbnb in Singapore are likely clustered. The p-value of 0.01 is smaller than 0.05. Hence, we reject the null hypothesis that the airbnb point patterns in Singapore are randomly distributed.

5.2.1.2.2 Clark and Evans Test: Kallang
  1. The following code chunk performs Clark-Evans test of aggregation for airbnb in Kallang using clarkevans.test() of spatstat
  2. The clipping region is set to NULL
  3. The type of alternative for the hypothesis is two.sided
  4. The number of simulations is set to 999
Show code 
clarkevans.test(airbnb0619_kl_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_kl_ppp
R = 0.57597, p-value = 0.002
alternative hypothesis: two-sided

Analysis: The p-value of 0.002 is smaller than 0.05. Hence, we reject the null hypothesis that the airbnb point patterns in Kallang are randomly distributed.

5.2.1.2.3 Clark and Evans Test: River Valley
  1. The following code chunk performs Clark-Evans test of aggregation for airbnb in River Valley using clarkevans.test() of spatstat
  2. The clipping region is set to NULL
  3. The type of alternative for the hypothesis is two.sided
  4. The number of simulations is set to 999
Show code 
clarkevans.test(airbnb0619_rv_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_rv_ppp
R = 0.83233, p-value = 0.002
alternative hypothesis: two-sided

Analysis: The p-value of 0.002 is smaller than 0.05. Hence, we reject the null hypothesis that the airbnb point patterns in River Valley are randomly distributed.

5.2.1.2.4 Clark and Evans Test: Novena
  1. The following code chunk performs Clark-Evans test of aggregation for airbnb in Novena using clarkevans.test() of spatstat
  2. The clipping region is set to NULL
  3. The type of alternative for the hypothesis is two.sided
  4. The number of simulations is set to 999
Show code 
clarkevans.test(airbnb0619_nv_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_nv_ppp
R = 0.55199, p-value = 0.002
alternative hypothesis: two-sided

Analysis: The p-value of 0.002 is smaller than 0.05. Hence, we reject the null hypothesis that the airbnb point patterns in Novena are randomly distributed.

5.2.1.2.5 Clark and Evans Test: Rochor
  1. The following code chunk performs Clark-Evans test of aggregation for airbnb in Rochor using clarkevans.test() of spatstat
  2. The clipping region is set to NULL
  3. The type of alternative for the hypothesis is two.sided
  4. The number of simulations is set to 999
Show code 
clarkevans.test(airbnb0619_rc_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_rc_ppp
R = 0.87405, p-value = 0.002
alternative hypothesis: two-sided

Analysis: The p-value of 0.002 is smaller than 0.05. Hence, we reject the null hypothesis that the airbnb point patterns in Rochor are randomly distributed.

Overall, we can conclude that airbnb in the four areas are not randomly distributed. For further analysis, we will narrow down to two areas, River valley and Rochor because they have R = 0.84 and 0.85 respectively, which is near 1. When R is nearer to 1, the distribution are likely random. Hence, we are interested to further analyse the distribution of airbnb within these two region.

5.2.2 G-FUNCTION

The G function measures the distribution of the distances from an arbitrary event to its nearest event.

River Valley

Compute G-Function Estimation

The following code chunk computes G-function using Gest() of spatat package.

Show code 
G_RV = Gest(airbnb0619_rv_ppp, correction = "border")
plot(G_RV, xlim=c(0,150))
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in River Valley are randomly distributed.

H1= The distribution of airbnb in River Valley are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs monta carlo simulation test using envelope() of spatstat package.

Show code 
G_RV.csr <- envelope(airbnb0619_rv_ppp, Gest, nsim = 999)

The following code chunk plots the csr

Show code 
plot(G_RV.csr)

Rochor

Compute G-Function Estimation

The following code chunk computes G-function using Gest() of spatat package.

Show code 
G_RC = Gest(airbnb0619_rc_ppp, correction = "border")
plot(G_RC, xlim=c(0,150))
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in Rochor are randomly distributed.

H1= The distribution of airbnb in Rochor are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs monta carlo simulation test using envelope() of spatstat package.

Show code 
G_RC.csr <- envelope(airbnb0619_rc_ppp, Gest, nsim = 999)

The following code chunk plots the csr

Show code 
plot(G_RC.csr)

F-FUNCTION

The F function estimates the empty space function F(r) or its hazard rate h(r) from a point pattern in a window of arbitrary shape.

River Valley

Compute F-Function Estimation

The following code chunk compute F-function using Fest() of spatat package.

Show code 
F_RV = Fest(airbnb0619_rv_ppp)
plot(F_RV)
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in River Valley are randomly distributed.

H1= The distribution of airbnb in River Valley are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs monta carlo simulation test using envelope() of spatstat package.

Show code 
F_RV.csr <- envelope(airbnb0619_rv_ppp, Fest, nsim = 999)

The following code chunk plots the csr

Show code 
plot(F_RV.csr)

Rochor

Compute F-Function Estimation

The following code chunk compute F-function using Fest() of spatat package.

Show code 
F_RC = Fest(airbnb0619_rc_ppp)
plot(F_RC)
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in Rochor are randomly distributed.

H1= The distribution of airbnb in Rochor are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs monta carlo simulation test using envelope() of spatstat package.

Show code 
F_RC.csr <- envelope(airbnb0619_rc_ppp, Fest, nsim = 999)

The following code chunk plots the csr

Show code 
plot(F_RC.csr)

K-FUNCTION

K-function measures the number of events found up to a given distance of any particular event.

River Valley

Compute K-Function Estimation

The following code chunk compute K-function using Kest() of spatat package.

Show code 
K_RV = Kest(airbnb0619_rv_ppp, correction = "Ripley")
plot(K_RV, . -r ~ r, ylab= "K(d)-r", xlab = "d(m)")
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in River Valley are randomly distributed.

H1= The distribution of airbnb in River Valley are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs the test using envelope() of spatstat package.

Show code 
K_RV.csr <- envelope(airbnb0619_rv_ppp, Kest, nsim = 99, rank = 1, glocal=TRUE)

The following code chunk plots the csr

Show code 
plot(K_RV.csr, . - r ~ r, xlab="d", ylab="K(d)-r")

Rochor

Compute K-Function Estimation

The following code chunk compute K-function using Kest() of spatat package.

Show code 
K_RC = Kest(airbnb0619_rc_ppp, correction = "Ripley")
plot(K_RC, . -r ~ r, ylab= "K(d)-r", xlab = "d(m)")
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in Rochor are randomly distributed.

H1= The distribution of airbnb in Rochor are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs the test using envelope() of spatstat package.

Show code 
K_RC.csr <- envelope(airbnb0619_rc_ppp, Kest, nsim = 99, rank = 1, glocal=TRUE)

The following code chunk plots the csr

Show code 
plot(K_RC.csr, . - r ~ r, xlab="d", ylab="K(d)-r")

L-FUNCTION

River Valley

Compute L-Function Estimation

The following code chunk compute L-function using Lest() of spatat package.

Show code 
L_RV = Lest(airbnb0619_rv_ppp, correction = "Ripley")
plot(L_RV, . -r ~ r, ylab= "L(d)-r", xlab = "d(m)")
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in River Valley are randomly distributed.

H1= The distribution of airbnb in River Valley are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs the test using envelope() of spatstat package.

Show code 
L_RV.csr <- envelope(airbnb0619_rv_ppp, Lest, nsim = 99, rank = 1, glocal=TRUE)

The following code chunk plots the csr

Show code 
plot(L_RV.csr, . - r ~ r, xlab="d", ylab="L(d)-r")

Rochor

Compute L-Function Estimation

The following code chunk compute K-function using Lest() of spatat package.

Show code 
L_RC = Lest(airbnb0619_rc_ppp, correction = "Ripley")
plot(L_RC, . -r ~ r, ylab= "L(d)-r", xlab = "d(m)")
Complete Spatial Randomness Test

Hypothesis:

Ho = The distribution of airbnb in Rochor are randomly distributed.

H1= The distribution of airbnb in Rochor are not randomly distributed.

The null hypothesis will be rejected is p-value is smaller than alpha value of 0.001.

The following code chunk performs the test using envelope() of spatstat package.

Show code 
L_RC.csr <- envelope(airbnb0619_rc_ppp, Lest, nsim = 99, rank = 1, glocal=TRUE)

The following code chunk plots the csr

Show code 
plot(L_RC.csr, . - r ~ r, xlab="d", ylab="L(d)-r")

6 Section B: Impact of COVID-19

6.1 Exploratory Spatial Data Analysis

6.1.1 Kernel density maps

6.1.1.1 Airbnb0721 listings

6.1.1.2.1 Rescale KDE values

The following code chunk rescale the KDE value of airbnb0721SG_ppp using rescale function of spatstat from meter to kilometer.

Show code 
airbnb0721SG_ppp.km <- rescale(airbnb0721SG_ppp, 1000, "km")
6.1.1.2.2 KDE Map
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable in km.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
  5. plot function of base R package to plot the kde map
Show code 
kde_airbnb0721SG_bw.km <- density(airbnb0721SG_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

plot(kde_airbnb0721SG_bw.km, main = "KDE Map")

6.1.1.2.3 Convert KDE output into grid object.

The following code chunk converts KDE output into grid object using as.SpatialGridDataFrame.im function of maptools, and plot the grid object using spplot function of raster package.

Show code 
gridded_kde_airbnb0721SG_bw <- as.SpatialGridDataFrame.im(kde_airbnb0721SG_bw.km)
spplot(gridded_kde_airbnb0721SG_bw)

6.1.1.2.4 Convert grid object into raster and assign project system

The following code chunk converts grid object into raster using raster function of raster package

Show code 
kde_airbnb0721SG_bw_raster <- raster(gridded_kde_airbnb0721SG_bw)
projection(kde_airbnb0721SG_bw_raster) <- CRS("+init=EPSG:3414")
kde_airbnb0721SG_bw_raster

class      : RasterLayer 
dimensions : 128, 128, 16384  (nrow, ncol, ncell)
resolution : 0.4170614, 0.2647348  (x, y)
extent     : 2.663926, 56.04779, 16.35798, 50.24403  (xmin, xmax, ymin, ymax)
crs        : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +units=m +no_defs 
source     : memory
names      : v 
values     : -2.02767e-13, 1050.623  (min, max)

6.1.1.2 Comparison between Airbnb0619 and Airbnb0721

6.1.1.2.1 KDE Map Comparison
Show code 
par(mfrow=c(1,2))
plot(kde_airbnb0619SG_bw.km, main = "Airbnb 2019")
plot(kde_airbnb0721SG_bw.km, main = "Airbnb 2021")

From the KDE map, the south-east region still have higher density in 2021. However, we notice the density of Airbnb have reduced as seen from the scale. It reduced from range 0-1500 in 2019 to range 0-1000 in 2021. We can conclude that Covid-19 did impact the Airbnb listing in Singapore from 2019 to 2021.

6.1.1.3 Comparison between Airbnb0619 and Airbnb0721 by Room Type

6.1.1.3.1 Data Wrangling
6.1.1.3.1.1 Determine the unique room type

The following code checks the unique room type using unique features of base R

Show code 
unique(airbnb0619_sf$room_type)

[1] "Private room"    "Entire home/apt" "Shared room"

We have determined that the unique room types are Private room, Entire home/apt and Shared room for Airbnb 2021.

Show code 
unique(airbnb0721_sf$room_type)

[1] "Private room"    "Entire home/apt" "Shared room"    
[4] "Hotel room"

We have determined that the unique room types are Private room, Entire home/apt, Shared room and Hotel room for Airbnb 2021.

6.1.1.3.1.2 Filter each room type from SpatialPointsData Frama
Show code 
airbnb0619_pte = airbnb0619[airbnb0619@data$room_type == "Private room",]
airbnb0619_ent = airbnb0619[airbnb0619@data$room_type == "Entire home/apt",]
airbnb0619_sh = airbnb0619[airbnb0619@data$room_type == "Shared room",]
airbnb0721_pte = airbnb0721[airbnb0721@data$room_type == "Private room",]
airbnb0721_ent = airbnb0721[airbnb0721@data$room_type == "Entire home/apt",]
airbnb0721_sh = airbnb0721[airbnb0721@data$room_type == "Shared room",]
airbnb0721_ht = airbnb0721[airbnb0721@data$room_type == "Hotel room",]
Show code 
airbnb0721_sh

class       : SpatialPointsDataFrame 
features    : 172 
extent      : 12801.45, 42673.57, 28189.47, 47475.94  (xmin, xmax, ymin, ymax)
crs         : +proj=tmerc +lat_0=1.36666666666667 +lon_0=103.833333333333 +k=1 +x_0=28001.642 +y_0=38744.572 +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs 
variables   : 14
names       :       id,                                            name,   host_id,                                            host_name, neighbourhood_group, neighbourhood,   room_type, price, minimum_nights, number_of_reviews, last_review, reviews_per_month, calculated_host_listings_count, availability_365 
min values  :  1277919, (Close to MRT) Single Mixed Dorm with Breakfast,    581033, <U+8BA4><U+8BC1><U+624B><U+673A><U+7528><U+6237>9873,      Central Region,    Ang Mo Kio, Shared room,    15,              1,                 0,       16618,              0.01,                              1,                0 
max values  : 50989011,           Your Private Pod Outside The City (2), 374499296,                                                 Zuzu,         West Region,        Yishun, Shared room,  3770,            240,               150,       18793,              2.11,                            165,              365

Notice that the data are in SpatialPointsData Frame format

6.1.1.3.1.3 Converting the Spatial* class into generic sp format

Next, we will convert Spatial* classes into Spatial object so that we can convert it into ppp object.

The following code chunk converts Spatial* classes into generic sp format using as function which sets the classes to their respective sp format.

Show code 
airbnb0619_pte_sp <- as(airbnb0619_pte, "SpatialPoints")
airbnb0619_ent_sp <- as(airbnb0619_ent, "SpatialPoints")
airbnb0619_sh_sp <- as(airbnb0619_sh, "SpatialPoints")
airbnb0721_pte_sp <- as(airbnb0721_pte, "SpatialPoints")
airbnb0721_ent_sp <- as(airbnb0721_ent, "SpatialPoints")
airbnb0721_sh_sp <- as(airbnb0721_sh, "SpatialPoints")
airbnb0721_ht_sp <- as(airbnb0721_ht, "SpatialPoints")
6.1.1.3.1.4 Converting the generic sp format into spatstat’s ppp format

We will next convert the spatial data into spatstat’s ppp format.

The following code chunk converts spatial data into spatstat’s ppp format using as function.

Show code 
airbnb0619_pte_ppp <- as(airbnb0619_pte_sp, "ppp")
airbnb0619_ent_ppp <- as(airbnb0619_ent_sp, "ppp")
airbnb0619_sh_ppp <- as(airbnb0619_sh_sp, "ppp")
airbnb0721_pte_ppp <- as(airbnb0721_pte_sp, "ppp")
airbnb0721_ent_ppp <- as(airbnb0721_ent_sp, "ppp")
airbnb0721_sh_ppp <- as(airbnb0721_sh_sp, "ppp")
airbnb0721_ht_ppp <- as(airbnb0721_ht_sp, "ppp")
6.1.1.3.1.5 Check for duplicates in each ppp’s objects

The following code chunk use any and duplicated function of base R package to check for duplicates, and sum up the count of duplicated points using sum function of base R package and multiplicity function of spatstat package.

Show code 
any(duplicated(airbnb0619_pte_ppp))

[1] FALSE
Show code 
sum(multiplicity(airbnb0619_pte_ppp) > 1)

[1] 0
Show code 
any(duplicated(airbnb0619_ent_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0619_ent_ppp) > 1)

[1] 2
Show code 
any(duplicated(airbnb0619_sh_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0619_sh_ppp) > 1)

[1] 2
Show code 
any(duplicated(airbnb0721_pte_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0721_pte_ppp) > 1)

[1] 126
Show code 
any(duplicated(airbnb0721_ent_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0721_ent_ppp) > 1)

[1] 40
Show code 
any(duplicated(airbnb0721_sh_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0721_sh_ppp) > 1)

[1] 16
Show code 
any(duplicated(airbnb0721_ht_ppp))

[1] TRUE
Show code 
sum(multiplicity(airbnb0721_ht_ppp) > 1)

[1] 35
6.1.1.3.1.6 Handling Duplicates

We will use jittering to handle duplicates so that we will not lose useful points. Jittering adds a small perturbation to the duplicated points so that they do not occupy the exact same space.

The following code chunk uses rjitter function of spatstat to jitter the ppp object with duplicates.

Show code 
airbnb0619_ent_ppp_jit <- rjitter(airbnb0619_ent_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

airbnb0619_sh_ppp_jit <- rjitter(airbnb0619_sh_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

airbnb0721_pte_ppp_jit <- rjitter(airbnb0721_pte_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

airbnb0721_ent_ppp_jit <- rjitter(airbnb0721_ent_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

airbnb0721_sh_ppp_jit <- rjitter(airbnb0721_sh_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)

airbnb0721_ht_ppp_jit <- rjitter(airbnb0721_ht_ppp, 
                             retry=TRUE, 
                             nsim=1, 
                             drop=TRUE)
6.1.1.3.1.7 Check if duplicates still exist
Show code 
any(duplicated(airbnb0619_ent_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0619_ent_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(airbnb0619_sh_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0619_sh_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(airbnb0721_pte_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0721_pte_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(airbnb0721_ent_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0721_ent_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(airbnb0721_sh_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0721_sh_ppp_jit) > 1)

[1] 0
Show code 
any(duplicated(airbnb0721_ht_ppp_jit))

[1] FALSE
Show code 
sum(multiplicity(airbnb0721_ht_ppp_jit) > 1)

[1] 0

We can confirm that there are no duplicated points after jittering.

6.1.1.3.1.8 Combining point events object and owin object

The following code chunk combines the point objects with the owin object into one ppp object class.

Show code 
airbnb0619SG_pte_ppp = airbnb0619_pte_ppp[sg_owin]
airbnb0619SG_ent_ppp = airbnb0619_ent_ppp_jit[sg_owin]
airbnb0619SG_sh_ppp = airbnb0619_sh_ppp_jit[sg_owin]
airbnb0721SG_pte_ppp = airbnb0721_pte_ppp_jit[sg_owin]
airbnb0721SG_ent_ppp = airbnb0721_ent_ppp_jit[sg_owin]
airbnb0721SG_sh_ppp = airbnb0721_sh_ppp_jit[sg_owin]
airbnb0721SG_ht_ppp = airbnb0721_ht_ppp_jit[sg_owin]
6.1.1.3.1.9 Rescale KDE values

The following code chunk rescale the KDE value of mrtlrtSG_ppp using rescale function of spatstat from meter to kilometer.

Show code 
airbnb0619SG_pte_ppp.km <- rescale(airbnb0619SG_pte_ppp, 1000, "km")
airbnb0619SG_ent_ppp.km <- rescale(airbnb0619SG_ent_ppp, 1000, "km")
airbnb0619SG_sh_ppp.km <- rescale(airbnb0619SG_sh_ppp, 1000, "km")
airbnb0721SG_pte_ppp.km <- rescale(airbnb0721SG_pte_ppp, 1000, "km")
airbnb0721SG_ent_ppp.km <- rescale(airbnb0721SG_ent_ppp, 1000, "km")
airbnb0721SG_sh_ppp.km <- rescale(airbnb0721SG_sh_ppp, 1000, "km")
airbnb0721SG_ht_ppp.km <- rescale(airbnb0721SG_ht_ppp, 1000, "km")
6.1.1.3.1.9 KDE
  1. The following code chunk computes kernel density using density function of spatstat package.
  2. The code first takes in the ppp object variable in km.
  3. bw.diggle automatic bandwidth method to detect a tight cluster
  4. smoothing kernel is set to default gaussian method
Show code 
kde_airbnb0619SG_pte_bw.km <- density(airbnb0619SG_pte_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

kde_airbnb0619SG_ent_bw.km <- density(airbnb0619SG_ent_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

kde_airbnb0619SG_sh_bw.km <- density(airbnb0619SG_sh_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

kde_airbnb0721SG_pte_bw.km <- density(airbnb0721SG_pte_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

kde_airbnb0721SG_ent_bw.km <- density(airbnb0721SG_ent_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

kde_airbnb0721SG_sh_bw.km <- density(airbnb0721SG_sh_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")

kde_airbnb0721SG_ht_bw.km <- density(airbnb0721SG_ht_ppp.km,
                              sigma=bw.diggle,
                              edge=TRUE,
                            kernel="gaussian")
6.1.1.3.1.10 Convert KDE output into grid object

The following code chunk converts KDE output into grid object using as.SpatialGridDataFrame.im function of maptools

Show code 
gridded_kde_airbnb0619SG_pte_bw <- as.SpatialGridDataFrame.im(kde_airbnb0619SG_pte_bw.km)
gridded_kde_airbnb0619SG_ent_bw <- as.SpatialGridDataFrame.im(kde_airbnb0619SG_ent_bw.km)
gridded_kde_airbnb0619SG_sh_bw <- as.SpatialGridDataFrame.im(kde_airbnb0619SG_sh_bw.km)
gridded_kde_airbnb0721SG_pte_bw <- as.SpatialGridDataFrame.im(kde_airbnb0721SG_pte_bw.km)
gridded_kde_airbnb0721SG_ent_bw <- as.SpatialGridDataFrame.im(kde_airbnb0721SG_ent_bw.km)
gridded_kde_airbnb0721SG_sh_bw <- as.SpatialGridDataFrame.im(kde_airbnb0721SG_sh_bw.km)
gridded_kde_airbnb0721SG_ht_bw <- as.SpatialGridDataFrame.im(kde_airbnb0721SG_ht_bw.km)
6.1.1.3.1.11 Convert grid object into raster and assign project system

The following code chunk converts grid object into raster using raster function of raster package

Show code 
kde_airbnb0619SG_pte_bw_raster <- raster(gridded_kde_airbnb0619SG_pte_bw)
projection(kde_airbnb0619SG_pte_bw_raster) <- CRS("+init=EPSG:3414")

kde_airbnb0619SG_ent_bw_raster <- raster(gridded_kde_airbnb0619SG_ent_bw)
projection(kde_airbnb0619SG_ent_bw_raster) <- CRS("+init=EPSG:3414")

kde_airbnb0619SG_sh_bw_raster <- raster(gridded_kde_airbnb0619SG_sh_bw)
projection(kde_airbnb0619SG_sh_bw_raster) <- CRS("+init=EPSG:3414")

kde_airbnb0721SG_pte_bw_raster <- raster(gridded_kde_airbnb0721SG_pte_bw)
projection(kde_airbnb0721SG_pte_bw_raster) <- CRS("+init=EPSG:3414")

kde_airbnb0721SG_ent_bw_raster <- raster(gridded_kde_airbnb0721SG_ent_bw)
projection(kde_airbnb0721SG_ent_bw_raster) <- CRS("+init=EPSG:3414")

kde_airbnb0721SG_sh_bw_raster <- raster(gridded_kde_airbnb0721SG_sh_bw)
projection(kde_airbnb0721SG_sh_bw_raster) <- CRS("+init=EPSG:3414")

kde_airbnb0721SG_ht_bw_raster <- raster(gridded_kde_airbnb0721SG_ht_bw)
projection(kde_airbnb0721SG_ht_bw_raster) <- CRS("+init=EPSG:3414")
6.1.1.3.2 KDE Map Comparison by Room type and Year

The following code chunks plot KDE map of Airbnb 2019 and 2021 by each room types side by side.

6.1.1.3.2.1 Private Room
Show code 
par(mfrow=c(1,2))
plot(kde_airbnb0619SG_pte_bw.km, main = "Airbnb 2019 Private Room")
plot(kde_airbnb0721SG_pte_bw.km, main = "Airbnb 2021 Private Room")

Analysis: There is a slight decrease in private room airbnb from 2019 to 2021 as seen in 2021 that there are lesser red point. The scale of 2019 to 2021 descreased slightly.

6.1.1.3.2.2 Entire Home/Apt
Show code 
par(mfrow=c(1,2))
plot(kde_airbnb0619SG_ent_bw.km, main = "Airbnb 2019 Entire Home/Apt")
plot(kde_airbnb0721SG_ent_bw.km, main = "Airbnb 2021 Entire Home/Apt")

Analysis: The map shows high density of airbnb in south-east region near central for both 2019 and 2021 as seen from the red marks on both map. However, there is a huge decrease in the number of Entire Home/Apt airbnb from 2019 to 2021 as seen on the scale. It decreased from 0-1000 in 2019 to 0-250 in 2021.

6.1.1.3.2.3 Shared Room
Show code 
par(mfrow=c(1,2))
plot(kde_airbnb0619SG_sh_bw.km, main = "Airbnb 2019 Shared Room")
plot(kde_airbnb0721SG_sh_bw.km, main = "Airbnb 2021 Shared Room")

Analysis: The map shows high density of airbnb in south-east region near central for both 2019 and 2021 as seen from the red marks on both map. However, there is a significant decrease in the number of Shared Room from 2019 to 2021 as seen on the scale. It ranges from 0-400 in 2019 and decrease to 0-80 in 2021.

6.1.1.3.2.4 Hotel Room
Show code 
plot(kde_airbnb0721SG_ht_bw.km, main = "Airbnb 2021 Hotel Room")

Analyis: Hotel room is a new room type in 2021.

Overall, we can conclude that Covid-19 are likely to have impact on the type of Airbnb room listed in Singapore from 2019 to 2021. It is seen from the significant decrease in Entire Home/Apt and Shared Room type, and this is likely because the owner do not wish to rent out the house anymore, or that there are lesser to no tourist. In addition, there is a new room type which is the Hotel Room, and this because hotel are required for quarantine purpose.

6.1.1.3.3 Exploration with Openstreetmap
Show code 
tmap_mode("view")

airbnb0619map <- tm_shape(airbnb0619_sf) +
                  tm_dots(col = 'room_type', palette=c(entire='blue', private='red', shared='green'), size = 0.02) +
                  tm_basemap(leaflet::providers$OpenStreetMap)

airbnb0721map <- tm_shape(airbnb0721_sf) +
                  tm_dots(col = 'room_type', palette=c(entire='blue', hotel='yellow', private='red', shared='green'), size = 0.02) + 
                  tm_basemap(leaflet::providers$OpenStreetMap)

tmap_arrange(airbnb0619map, airbnb0721map)
Show code 
tmap_mode("plot")

6.2 Second-order Spatial Point Patterns Analysis

6.2.1 Nearest Neighbour Analysis (CONTINUE FROM HERE)

6.2.1.1 Hypothesis

we will perform the Clark-Evans test of aggregation for a spatial point pattern by using clarkevans.test() of statspat.

Test Hypothesis:

H0 = The distribution of airbnb are randomly distributed.

H1 = The distribution of airbnb are not randomly distributed.

The 95% confident interval will be used.

6.2.1.2 Testing and Analysis

6.2.1.2.1 Testing spatial point patterns using Clark and Evans Test
Show code 
clarkevans.test(airbnb0619SG_ppp,
                correction="none",
                clipregion="sg_owin",
                alternative=c("clustered"),
                nsim=99)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 99 simulations of CSR with fixed n

data:  airbnb0619SG_ppp
R = 0.36024, p-value = 0.01
alternative hypothesis: clustered (R < 1)

Clark-Evans test No edge correction Monte Carlo test based on 99 simulations of CSR with fixed n

data: airbnb0619SG_ppp R = 0.36169, p-value = 0.01 alternative hypothesis: clustered (R < 1)

5.2.1.2.2 Clark and Evans Test: Kallang planning area
Show code 
clarkevans.test(airbnb0619_kl_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_kl_ppp
R = 0.57597, p-value = 0.002
alternative hypothesis: two-sided

Analysis:

5.2.1.2.3 Clark and Evans Test: River Valley planning area
Show code 
clarkevans.test(airbnb0619_rv_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_rv_ppp
R = 0.83233, p-value = 0.002
alternative hypothesis: two-sided

Analysis:

5.2.1.2.4 Clark and Evans Test: Novena planning area
Show code 
clarkevans.test(airbnb0619_nv_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_nv_ppp
R = 0.55199, p-value = 0.002
alternative hypothesis: two-sided

Analysis:

5.2.1.2.5 Clark and Evans Test: Rochor planning area
Show code 
clarkevans.test(airbnb0619_rc_ppp,
                correction="none",
                clipregion=NULL,
                alternative=c("two.sided"),
                nsim=999)


    Clark-Evans test
    No edge correction
    Monte Carlo test based on 999 simulations of CSR with fixed n

data:  airbnb0619_rc_ppp
R = 0.87405, p-value = 0.002
alternative hypothesis: two-sided

Analysis:

G-FUNCTION

Hypothesis

Testing and Analysis

F-FUNCTION

Hypothesis

Testing and Analysis

K-FUNCTION

Hypothesis

Testing and Analysis

L-FUNCTION

Hypothesis

Testing and Analysis